On 2/2/16 6:39 PM, Tom Lane wrote:
> I'm inclined to think that a good solution would be to create an
> artificial restriction to not accept years beyond, say, 100000 AD.
> That would leave us with a lot of daylight to not have to worry
> about corner-case overflows in timestamp arithmetic. I'm not sure
> though where we'd need to enforce such a restriction; certainly in
> timestamp[tz]_in, but where else?
Probably some of the casts (I'd think at least timestamp->timestamptz).
Maybe timestamp[tz]_recv. Most of the time*pl* functions. :/
--
Jim Nasby, Data Architect, Blue Treble Consulting, Austin TX
Experts in Analytics, Data Architecture and PostgreSQL
Data in Trouble? Get it in Treble! http://BlueTreble.com